Landmark feature determination method and device, electronic equipment and storage medium

Abstract

The present disclosure provides a landmark feature determination method and device, electronic equipment and storage medium. The landmark feature determination method comprises: obtaining target expression data, wherein the target expression data is used to represent the expression amount of each candidate feature in different cells; the average expression amount and expression proportion of each candidate feature in different cell types are counted; for each candidate feature, the first inter-class comparison data is determined according to the average expression amount of the candidate feature in the target cell type and the average expression amount of the candidate feature in the non-target cell type; for each candidate feature, the second inter-class comparison data is determined according to the expression proportion of the candidate feature in the target cell type and the expression proportion of the candidate feature in the non-target cell type; and the landmark feature of the target cell type is determined according to the first inter-class comparison data and the second inter-class comparison data. The embodiment of the present application can solve the problem of landmark feature specificity.

Description

Technical Field

[0001] This disclosure relates to the field of bioinformatics, specifically to a method and apparatus for determining biomarkers, electronic equipment, and storage medium. Background Technology

[0002] Currently, different cell types can be distinguished by marker features, which can include marker genes (i.e., marker genes) and marker regions (i.e., marker peaks).

[0003] Taking marker genes as an example, related technologies identify marker genes by comparing the average gene expression levels in cell populations or tissues. However, this method is prone to low cell type specificity for marker genes, meaning that a definite marker gene may be highly expressed in multiple cell types. Therefore, how to provide a marker feature determination method that can solve the problem of marker feature specificity has become an urgent technical problem to be solved. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, electronic device, and storage medium for determining marker features, aiming to solve the problem of marker feature specificity.

[0005] To achieve the above objectives, a first aspect of this application proposes a method for determining a marker feature, the method comprising:

[0006] Obtain target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells;

[0007] The average expression level and expression ratio of each candidate feature in different cell types were statistically analyzed.

[0008] For each candidate feature, inter-class comparison data is determined based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type.

[0009] For each candidate feature, inter-class comparison data is determined based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type.

[0010] The marker features of the target cell type are determined based on the first inter-class comparison data and the second inter-class comparison data.

[0011] In some embodiments, determining the first-class comparison data for each candidate feature based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in non-target cell types includes:

[0012] The average expression levels of the candidate features in multiple non-target cell types are compared to determine the reference average expression level with the largest value.

[0013] The average expression difference is calculated by comparing the average expression level of the candidate feature in the target cell type with the reference average expression level.

[0014] The first inter-class comparison data is determined from the average expression difference and the first preset value.

[0015] In some embodiments, determining the second-class comparison data for each candidate feature based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type includes:

[0016] The expression ratios of the candidate features in multiple non-target cell types are compared to determine the reference expression ratio with the largest value.

[0017] The difference between the expression ratio of the candidate feature in the target cell type and the reference expression ratio is calculated to obtain the ratio difference.

[0018] The second inter-class comparison data is determined from the ratio difference and the second preset value.

[0019] In some embodiments, determining the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data includes:

[0020] Determine the first comparison threshold and the second comparison threshold;

[0021] If the first inter-class comparison data is greater than the first comparison threshold, and the second inter-class comparison data is greater than the second comparison threshold, then the candidate feature is used as the marker feature.

[0022] In some embodiments, after determining the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data, the method further includes:

[0023] Feature relationship data is constructed based on the target cell type, the corresponding marker features of the target cell type, and the candidate features;

[0024] The first comparison threshold and the second comparison threshold are adjusted based on the feature relationship data.

[0025] In some embodiments, a method for calculating the average expression level of each candidate feature across different cell types includes:

[0026] For each cell type, the target expression level of the candidate feature in that cell type is calculated.

[0027] Calculate the total expression level of the candidate features across all cell types;

[0028] For each cell type, the average expression level of the candidate feature in that cell type is calculated based on the target expression level and the total expression level corresponding to that cell type.

[0029] In some embodiments, a method for calculating the expression ratio of each candidate feature in different cell types includes:

[0030] For each cell type, the number of first cells in the cell type that express non-zero expression of the candidate feature is counted, and the number of second cells in the cell type is also counted.

[0031] For each cell type, the expression ratio of the candidate feature in that cell type is calculated based on the number of first cells and the number of second cells corresponding to that cell type.

[0032] To achieve the above objectives, a second aspect of this application provides a marker feature determination apparatus, the apparatus comprising:

[0033] A data acquisition unit is used to acquire target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells;

[0034] The expression level determination unit is used to calculate the average expression level and expression ratio of each candidate feature in different cell types.

[0035] The first inter-class comparison unit is used to determine inter-class comparison data for each candidate feature based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type.

[0036] The second inter-class comparison unit is used to determine the second inter-class comparison data for each candidate feature based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type.

[0037] The marker feature determination unit is used to determine the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data.

[0038] To achieve the above objectives, a third aspect of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0039] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0040] To achieve the above objectives, a fifth aspect of this application provides a computer program product comprising a computer program that is read and executed by a processor of a computer device, causing the computer device to perform the method described in the first aspect.

[0041] The marker feature determination method, apparatus, electronic device, and storage medium proposed in this application can determine whether a candidate feature is highly expressed in a certain cell type by using the average expression and expression ratio of each candidate feature in different cell types. By comparing the expression of the candidate feature in the target cell type with its expression in non-target cell types using first-class and second-class comparison data, it is possible to determine whether the expression of the candidate feature in the target cell type is significant compared to non-target cell types. Thus, when determining the marker feature of the target cell type based on the first-class and second-class comparison data, the possibility of the marker feature being highly expressed in multiple cell types can be reduced, thereby solving the problem of marker feature specificity. Attached Figure Description

[0042] The accompanying drawings are provided to further understand the technical solutions of this disclosure and constitute a part of the specification. They are used together with the embodiments of this disclosure to explain the technical solutions of this disclosure and do not constitute a limitation on the technical solutions of this disclosure.

[0043] Figure 1 This is a flowchart of the marker feature determination method provided in the embodiments of this application;

[0044] Figure 2 This is a flowchart illustrating the determination of the average expression level of each cell in different cell types, provided in an embodiment of this application.

[0045] Figure 3 This is a flowchart provided in an embodiment of the present application for determining the expression ratio of each candidate feature in different cell types;

[0046] Figure 4 yes Figure 1 The flowchart of step S103 in the process;

[0047] Figure 5 yes Figure 1 The flowchart of step S104 in the process;

[0048] Figure 6 yes Figure 1The flowchart of step S105 in the process;

[0049] Figure 7 This is a flowchart of another embodiment of the marker feature determination method provided in this application;

[0050] Figure 8 This is a schematic diagram of the feature relationship data provided in the embodiments of this application;

[0051] Figure 9 This is a schematic diagram of the marker feature determination device provided in the embodiments of this application;

[0052] Figure 10 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the scope of this disclosure.

[0054] Before providing a further detailed description of the embodiments of this disclosure, the terms and concepts used in these embodiments are explained, and they are subject to the following interpretations:

[0055] Marker genes are genes whose expression levels are significantly higher in a specific cell type than in other cell types. Marker genes serve as identifiers for cell types, helping to identify and differentiate different cell populations.

[0056] Marker peaks: In scATAC-seq (a single-cell genomics technique that combines transposase-accessible chromatin sequencing and single-cell sequencing to identify open chromatin regions in each cell), marker peaks represent regions of chromatin accessibility. Chromatin accessibility refers to the property of certain regions of DNA to be bound by proteins such as transcription factors. In single-cell ATAC-seq data, marker peaks for each cell reflect the characteristics of chromatin accessibility in that cell, which can be used to distinguish different cell types. By analyzing the read distribution within marker peaks, open chromatin regions associated with specific cell types can be identified.

[0057] Time complexity refers to the time required for an algorithm to execute. Time complexity is typically represented using the O notation, for example: O(1): constant time complexity, meaning the algorithm's execution time does not change with the size of the input data. O(n): linear time complexity, meaning the algorithm's execution time is proportional to the size of the input data.

[0058] Space complexity refers to the maximum memory space required by an algorithm during execution. Space complexity is often represented using the O notation, for example: O(1): constant space complexity, meaning the memory space required by the algorithm does not change with the size of the input data. O(n): linear space complexity, meaning the memory space required by the algorithm is proportional to the size of the input data.

[0059] Currently, different cell types can be distinguished using marker features, which can include marker genes and marker regions. Taking marker genes as an example, the technology for identifying marker genes has undergone a leap forward from traditional batch RNA sequencing and immunohistochemical labeling to today's single-cell RNA sequencing (scRNA-seq). Early on, marker gene identification relied on batch RNA sequencing technology, primarily using differential expression analysis to identify genes significantly expressed under specific conditions. Traditional differential expression analysis methods include t-tests (a statistical hypothesis test used to determine if there is a significant difference between two sample means) and ANOVA (analysis of variance, a statistical method used to test if there are significant differences between the means of three or more samples), typically identifying marker genes by comparing the average gene expression levels in cell populations or tissues. While these methods are effective, they struggle to handle complex cellular heterogeneity, especially for detecting rare cell populations and low-abundance genes.

[0060] With the rise of single-cell RNA sequencing technology, marker gene discovery has entered a new phase. scRNA-seq enables the study of gene expression differences at single-cell resolution, significantly improving the accuracy of marker gene screening. Early analytical methods relied on traditional statistical tests (such as t-tests and Wilcoxon tests, which are nonparametric statistical hypothesis tests used to determine if there is a significant difference between the distributions of two samples). However, these methods gradually revealed limitations in handling highly sparse single-cell data, as single-cell data typically contains a large number of zero values ​​and technical noise. To address these issues, new statistical methods have been introduced. For example, MAST (Model-Adjusted scRNA-seq) uses a generalized linear model to handle technical noise and zero-value data, demonstrating better robustness in differential expression analysis. Meanwhile, models based on the negative binomial distribution, such as DESeq2 and edgeR, initially applied to batch RNA sequencing, have also been adapted to single-cell data, enabling the handling of gene expression data with a high dynamic range.

[0061] Currently, Bayesian methods (such as SCDE) further improve the ability to handle sparse data by modeling with a zero-inflated negative binomial distribution. With increasing data complexity, machine learning and deep learning methods are also gradually being introduced into marker gene identification, helping to automatically extract specifically expressed genes from large amounts of high-dimensional data. Furthermore, the emergence of new technologies such as spatial transcriptomics has promoted marker gene research that incorporates spatial information, enabling the identification of spatially specific marker genes within tissue structures.

[0062] In summary, the methods in related technologies mainly rely on comparing the mean expression of genes in different cell types or populations to identify marker genes. However, this method has the following drawbacks: 1. The specificity of marker genes for cell types is not high, and it usually includes some genes that are highly expressed in multiple cell types; 2. It is sensitive to noise; 3. It cannot efficiently process large-scale datasets; 4. Using multiple tests may introduce false positives; 5. It makes certain assumptions about the data distribution, which limits its applicability.

[0063] Based on this, embodiments of this application provide a method and apparatus for determining marker features, an electronic device, and a storage medium, which can solve the problem of the specificity of marker features.

[0064] The marker feature determination method provided in this application relates to the field of bioinformatics. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the marker feature determination method, but is not limited to the above forms.

[0065] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0066] The method for determining the marker features provided in the embodiments of this application will be described below.

[0067] Reference Figure 1 In some embodiments, the marker feature determination method provided in this application includes, but is not limited to, steps S101 to S105.

[0068] Step S101: Obtain target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells;

[0069] Step S102: Calculate the average expression level and expression ratio of each candidate feature in different cell types;

[0070] Step S103: For each candidate feature, determine the first inter-class comparison data based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type;

[0071] Step S104: For each candidate feature, determine the second inter-class comparison data based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type.

[0072] Step S105: Determine the marker features of the target cell type based on the comparison data between the first and second classes.

[0073] Steps S101 to S105, as illustrated in the embodiments of this application, determine whether a candidate feature is highly expressed in a certain cell type by using the average expression and expression ratio of each candidate feature in different cell types. By comparing the expression of the candidate feature in the target cell type with its expression in non-target cell types using first-class and second-class comparison data, it is possible to determine whether the expression of the candidate feature in the target cell type is significant compared to non-target cell types. Thus, when determining the marker feature of the target cell type based on the first-class and second-class comparison data, the possibility of the marker feature being highly expressed in multiple cell types can be reduced, thereby solving the problem of marker feature specificity.

[0074] In step S101 of some embodiments, candidate features can refer to elements capable of cell-specific representation, such as genes or chromatin accessibility regions. Target expression data can refer to data used to describe the relationship between multiple candidate features and different cells. Target expression data can be in matrix form, where the rows can represent different candidate features, the columns can represent different cells, and the elements corresponding to the rows and columns can indicate the expression level of the candidate feature in the cell. For example, assuming matrix X is the matrix corresponding to the target expression data, when the candidate feature is a gene, the element x in the m-th row and n-th column of matrix X... mn This can refer to the expression level of the gene corresponding to the m-th row (hereinafter referred to as gene m) in the cell corresponding to the n-th column (hereinafter referred to as cell n). Similarly, when the candidate feature is a chromatin accessibility region, the element x in the m-th row and n-th column of matrix X... mn This can refer to the expression level of chromatin-accessible region m in cell n. In other words, the expression level of chromatin-accessible region m in cell n can refer to the number of DNA fragments in cell n that fall within chromatin-accessible region m. Both m and n can be positive integers greater than zero. Cells can be identified from target slices or other methods (such as datasets). For each cell, the expression levels of different genes in that cell can be determined through operations such as RNA sequencing, or the expression levels of different chromatin-accessible regions in the cell can be determined through operations such as scATAC-seq.

[0075] In step S102 of some embodiments, the multiple cells in the target expression data may correspond to different cell types. For each cell type, the average expression level may refer to the ratio of the total expression level of a candidate feature in that cell type to the total expression level of that candidate feature in all cells. The expression ratio may refer to the proportion of cells in that cell type that have non-zero expression of a candidate feature.

[0076] Reference Figure 2In some embodiments, the method for calculating the average expression level of each candidate feature in different cell types includes, but is not limited to, steps S201 to S203.

[0077] Step S201: For each cell type, count the target expression level of the candidate feature in the cell type;

[0078] Step S202: Calculate the total expression level of candidate features across all cell types;

[0079] Step S203: For each cell type, calculate the average expression level of the candidate feature in the cell type based on the target expression level and total expression level corresponding to the cell type.

[0080] In step S201 of some embodiments, the cell type corresponding to each cell in the target expression data can be determined according to a clustering method. The clustering method can be adaptively set according to actual needs, and this embodiment of the application does not specifically limit it. For example, the clustering method can be the Leiden algorithm. The Leiden algorithm is a graph clustering algorithm that creates clusters by considering the ratio between the connection data between cells in the cluster and the connection data of the overall dataset. For each cell type, the target expression level can refer to the sum of the expression levels of a candidate feature in the cells corresponding to that cell type. Taking gene g1 as an example, assuming cell type A and cell type B. If cell type A includes cells C1, C2, and C3, then the target expression level of gene g1 in cell type A = the expression level of gene g1 in cell C1 + the expression level of gene g1 in cell C2 + the expression level of gene g1 in cell C3. If cell type B includes cells C4 and C5, then the target expression level of gene g1 in cell type B = the expression level of gene g1 in cell C4 + the expression level of gene g1 in cell C5.

[0081] In step S202 of some embodiments, for each candidate feature, the total expression level can refer to the sum of the expression levels of the candidate feature in all cells corresponding to all cell types. That is, the total expression level can refer to the sum of the expression levels of the candidate feature in all cells corresponding to the target expression data. Taking gene g1 as an example as the candidate feature, assuming that cell type C is also included, then the total expression level of gene g1 = the target expression level of gene g1 in cell type A + the target expression level of gene g1 in cell type B + the target expression level of gene g1 in cell type C.

[0082] In step S203 of some embodiments, for each cell type, the ratio of the target expression level to the total expression level for that cell type is calculated to obtain the average expression level of the candidate feature in that cell type. For example, the average expression level of the candidate feature in a certain cell type can be calculated according to the following formula (1).

[0083]

[0084] Among them, a mk This can represent the ratio (i.e., the target expression level) of the total expression level of candidate feature m in the k-th cell type (hereinafter referred to as cell type k, where k is the index of a certain cell type) to the total expression level of candidate feature m across all cell types. l∈c(k) x ml This can represent the target expression level of candidate feature m in the kth cell type, ∑ k∈[1,K] ∑ l∈c(k) x ml This can represent the total expression level of candidate feature m across all cell types. c(k) can be the index set of all cells belonging to cell type k. Let |c(k)| represent the size of cell type k (i.e., the total number of cells belonging to cell type k), and let [1,K] represent the set {1,2,...,K}, which can include K cell types.

[0085] Reference Figure 3 In some embodiments, the method for calculating the expression ratio of each candidate feature in different cell types may include steps S301 to S302.

[0086] Step S301: For each cell type, count the number of first cells in the cell type that have non-zero expression of candidate features, and count the number of second cells in the cell type.

[0087] Step S302: For each cell type, calculate the expression ratio of the candidate feature in the cell type based on the first cell number and the second cell number corresponding to the cell type.

[0088] In step S301 of some embodiments, non-zero expression can refer to an expression level that is not equal to 0. For each cell type, the first cell number can refer to the total number of cells in that cell type that have non-zero expression on a certain candidate feature. The second cell number can refer to the total number of all cells belonging to that cell type in the target expression data. Taking gene g1 as an example, assuming that cell type A includes cells C1, C2, and C3, the expression level of gene g1 in cell C1 is 0, and the expression level of gene g1 in cells C2 and C3 is greater than zero. Therefore, for gene g1, the first cell number of cell type A is 1, and the second cell number is 3.

[0089] In step S302 of some embodiments, for each cell type, the ratio of the first cell number and the second cell number corresponding to that cell type is calculated to obtain the expression ratio. For example, the expression ratio of the candidate feature in a certain cell type can be calculated according to the following formula (2).

[0090]

[0091] Among them, b mk This can represent the proportion (i.e., expression ratio) of cells in cell type k that have non-zero expression on candidate feature m. ∑ l∈c(k) φ(x ml Let |c(k)| represent the first number of cells of cell type k, and |c(k)| represent the second number of cells of cell type k. It is understandable that when x... ml When >1, φ(x) ml ) = 1; otherwise, φ(x) = 1. ml ) = 0.

[0092] The advantage of steps S301 to S302 is that by setting the expression ratio b related to non-zero expression... mk This variable can reduce the impact of high noise on the results of identifying marker features, thereby improving the accuracy of identifying marker features.

[0093] In step S103 of some embodiments, the target cell type may refer to the cell type of the marker feature to be determined in the target expression data. The non-target cell type may refer to other cell types in the target expression data besides the target cell type. For example, assuming cell type A, cell type B, and cell type C are included, when the target cell type is cell type A, the non-target cell types are cell type B and cell type C. When the target cell type is cell type B, the non-target cell types are cell type A and cell type C. The first inter-class comparison data may refer to data obtained by comparing the average expression level corresponding to the target cell type with the average expression level corresponding to the non-target cell type. Therefore, the first inter-class comparison data can reflect whether the candidate feature is significantly expressed in the target cell type.

[0094] Reference Figure 4 In some embodiments, step S103 may include, but is not limited to, steps S401 to S403.

[0095] Step S401: Compare the average expression levels of candidate features in multiple non-target cell types and determine the reference average expression level with the largest value.

[0096] Step S402: Calculate the difference between the average expression level of the candidate feature in the target cell type and the reference average expression level to obtain the average expression difference;

[0097] Step S403: Determine the first inter-class comparison data from the average expression difference and the first preset value.

[0098] In step S401 of some embodiments, for each candidate feature, the average expression level of the candidate feature in different non-target cell types is compared to determine the average expression level with the highest value in the non-target cell types. This highest average expression level is used as the reference average expression level. For example, taking cell type A as the target cell type and gene g1 as the candidate feature, the average expression level of gene g1 in cell type B is compared with the average expression level of gene g1 in cell type C. If the average expression level of gene g1 in cell type B is greater than the average expression level in cell type C, then the average expression level of cell type B is used as the reference average expression level.

[0099] In step S402 of some embodiments, for each candidate feature, the difference between the average expression level of the candidate feature in the target cell type and the reference average expression level corresponding to the candidate feature is calculated to obtain the average expression difference. For example, in the example of step S401, the difference between the average expression level of gene g1 in cell type A and the average expression level of gene g1 in cell type B can be calculated to obtain the average expression difference of gene g1 in cell type A.

[0100] In step S403 of some embodiments, the first preset value may refer to a pre-set value. The specific value of the first preset value can be adaptively set according to actual needs, and this application embodiment does not specifically limit this. The first preset value is compared with the average expression difference, and the value with the larger value is taken as the first inter-class comparison data. For example, the first inter-class comparison data of a certain candidate feature in the target cell type can be determined according to the following formula (3).

[0101]

[0102] Where, α mk This represents the first inter-class comparison data of candidate feature m in cell type k (i.e., cell type k is the target cell type). This represents the reference average expression level. It is understandable that in equation (3), setting the first preset value to 0 ensures that the inter-class comparison data is non-negative, thus allowing determination of the likelihood that candidate feature m belongs to a cell type k marker feature. For example, when the inter-class comparison data is 0, it indicates that the average expression difference is less than or equal to 0. This indicates that, compared to other cell types, the expression of candidate feature m in cell type k is not significant, and therefore, the likelihood that candidate feature m is a cell type k marker feature is low. Therefore, subsequent operations on candidate feature m can be omitted to reduce the waste of computational resources.

[0103] In step S104 of some embodiments, the second inter-class comparison data may refer to data obtained by comparing the expression ratio corresponding to the target cell type with the expression ratio corresponding to the non-target cell type. Therefore, the second inter-class comparison data can reflect whether the non-zero expression of the candidate feature in the target cell type is significant.

[0104] Reference Figure 5 In some embodiments, step S104 may include, but is not limited to, steps S501 to S503.

[0105] Step S501: Compare the expression ratios of candidate features in multiple non-target cell types and determine the reference expression ratio with the largest value;

[0106] Step S502: Calculate the difference between the expression ratio of the candidate feature in the target cell type and the reference expression ratio to obtain the ratio difference.

[0107] Step S503: Determine the second type of inter-class comparison data from the ratio difference and the second preset value.

[0108] In step S501 of some embodiments, for each candidate feature, the expression ratio of that candidate feature in multiple non-target cell types is compared to determine the expression ratio with the highest value among the non-target cell types. This highest value expression ratio is taken as the target expression ratio. For example, taking cell type A as the target cell type and gene g1 as the candidate feature, the expression ratio of gene g1 in cell type B is compared with the expression ratio of gene g1 in cell type C. If the expression ratio of gene g1 in cell type C is greater than that in cell type B, then the expression ratio of cell type C is taken as the reference expression ratio.

[0109] In step S502 of some embodiments, for each candidate feature, the difference between the expression ratio of the candidate feature in the target cell type and the expression ratio corresponding to the candidate feature is calculated to obtain the ratio difference. For example, in the example of step S501, the difference between the expression ratio of gene g1 in cell type A and the expression ratio of gene g1 in cell type C can be calculated to obtain the ratio difference of gene g1 in cell type A.

[0110] In step S503 of some embodiments, the second preset value may refer to a pre-set value. The specific value of the second preset value can be adaptively set according to actual needs, and this application embodiment does not specifically limit it. The second preset value is compared with the ratio difference, and the value with the larger value is taken as the second inter-class comparison data. For example, the second inter-class comparison data of a certain candidate feature in the target cell type can be determined according to the following formula (4).

[0111]

[0112] Where, β mk This represents the second-class comparison data of candidate feature m in cell type k (i.e., cell type k is the target cell type). The reference expression ratio is indicated. It is understood that in equation (4), setting the second preset value to 0 ensures that the inter-class comparison data is non-negative, which can then be used to determine the probability that candidate feature m belongs to the cell type k marker feature. For example, when the inter-class comparison data is 0, it indicates that the ratio difference is less than or equal to 0. This indicates that, compared to other cell types, the non-zero expression ratio of candidate feature m in cell type k is not significant, thus determining that the probability of candidate feature m being a cell type k marker feature is low. Therefore, subsequent operations on candidate feature m can be omitted to reduce the waste of computational resources.

[0113] In step S105 of some embodiments, for each candidate feature, the first inter-class comparison data and the second inter-class comparison data can reflect the difference between the expression of the candidate feature in the target cell type and its expression in non-cell types. Thus, when the expression of the corresponding candidate feature in the target cell type, determined based on the first inter-class comparison data and the second inter-class comparison data, differs significantly from its expression in non-cell types, it indicates that the candidate feature is highly expressed only in the target cell type. Therefore, the candidate feature can be used as a marker feature of the target cell type.

[0114] Reference Figure 6 In some embodiments, step S105 may include, but is not limited to, steps S601 to S602.

[0115] Step S601: Determine the first comparison threshold and the second comparison threshold;

[0116] Step S602: If the comparison data between the first class is greater than the first comparison threshold and the comparison data between the second class is greater than the second comparison threshold, then the candidate feature is used as the label feature.

[0117] In step S601 of some embodiments, the first comparison threshold may refer to a threshold pre-set for the comparison data between the first and second classes, and the second comparison threshold may refer to a threshold pre-set for the comparison data between the second and third classes. The specific values ​​of the first and second comparison thresholds can be adaptively set according to actual needs, and this embodiment does not specifically limit this. For example, the first and second comparison thresholds can be determined based on prior knowledge. Wherein, the first comparison threshold th... α It can be set to 0.5, the second comparison threshold th β It can be set to 0.6.

[0118] In step S602 of some embodiments, the first type comparison data α is... mk Comparison with the first threshold th α Compare the data and use the second-class comparison data β. mk Comparison with the second threshold th β Comparison. Compare data α between the first categories. mk Greater than the first comparison threshold th α And the second type of inter-class comparison data β mk Greater than the second comparison threshold th β In this case, it indicates that the candidate feature m is not only highly expressed in cell type k, but also highly expressed only in cell type k. Thus, the candidate feature m can be identified as a marker feature of cell type k (i.e., the target cell type). Therefore, the embodiments of this application can reduce the situation where a marker feature is highly expressed in multiple cell types.

[0119] Understandably, for each target cell type, the above operations are performed based on the expression levels of different candidate features in that target cell type, thus determining multiple marker features corresponding to each cell type. Based on these multiple marker features, a marker feature set S, as shown in equation (5), can be constructed. k .

[0120] S k ={m∈[1,M]|α mk >th α ,β mk >th β Equation (5)

[0121] Among them, S k Let m represent the set of features for cell type k, k∈[1,K]. Let m represent the candidate features, m∈[1,M], where M represents the total number of candidate features in the target expression data.

[0122] Reference Figure 7 In some embodiments, the marker feature determination method provided in this application may include, but is not limited to, steps S701 to S702.

[0123] Step S701: Construct feature relationship data based on the target cell type, the corresponding marker features, and candidate features;

[0124] Step S702: Adjust the first comparison threshold and the second comparison threshold according to the feature relationship data.

[0125] In step S701 of some embodiments, the feature relationship data can be used to describe the marker features corresponding to each target cell type among multiple target cell types and multiple candidate features. The feature relationship data can be in the form of a dot plot, for example... Figure 8 As shown. In Figure 8 In the table, row names represent different target cell types, and column names represent different candidate features. Figure 8 The color of the midpoint represents the average expression level of the candidate feature in the corresponding cell type; a redder midpoint indicates a higher average expression level. The size of the midpoint represents the proportion of cells expressing the candidate feature in that cell type; a larger midpoint indicates a higher proportion. Thus, we can... Figure 8 The red dots are considered as marker features. By comparing the marker features of different cell types, it can be determined whether there are duplicate marker features among different cell types. When duplicates or near-duplicates exist, the first and second comparison thresholds can be adjusted to improve the accuracy of marker feature determination.

[0126] The marker feature determination method provided in this application determines marker features through inter-class comparison data and inter-class comparison data. This method can suppress candidate features that are highly expressed in multiple cell types, allowing candidate features that are significantly expressed only in a specific cell type to be effectively identified. This solves the problem of marker feature nonspecificity caused by simultaneous high expression of candidate features in multiple cell types. Secondly, this application does not rely on hypothesis testing and p-value adjustment, thereby reducing various testing problems common in traditional differential expression analysis. This application makes no assumptions about data distribution, making it applicable to any single-cell and spatial single-cell transcriptome data. Furthermore, the marker feature can be any feature other than a marker gene (such as a marker peak in scATAC-seq data), therefore, this application is also applicable to other non-transcriptome data. Finally, compared to related technologies that determine the average expression of a gene in a target cell type and then determine the marker gene for that target cell type, this application determines the ratio of the total expression level of a gene in the target cell type to the total expression level of that gene in the entire dataset. Thus, the embodiments of this application can finally obtain a variable that varies in the range of 0 to 1, that is, the variation range of the first inter-class comparison data and the second inter-class comparison data is between 0 and 1, which makes it easier to determine the first comparison threshold and the second comparison threshold.

[0127] In a specific embodiment, taking a marker gene as an example, assuming the dataset includes N cells, M genes, and K cell types (or clusters), the method for determining the marker gene includes:

[0128] Step 1: For any cell type k∈[1,K] and any gene m∈[1,M], calculate the average expression level a based on the above equation (1). mk And based on the above formula (2), the expression ratio b is calculated. mk .

[0129] Step 2: For any gene m∈[1,M] in any cell type k∈[1,K], based on the average expression level a mk The first type of inter-class comparison data α is calculated using equation (3) above. mk Based on the expression ratio b mk The second type of inter-class comparison data β is calculated using equation (4) above. mk .

[0130] Step 3: For any cell type k∈[1,K] and any gene m∈[1,M], based on the first inter-class comparison data α mk Second type of inter-class comparison data β mk The set of marker genes S is determined by equation (5) above. k .

[0131] Understandably, the time complexity of the first step above is O(n). The time complexity of the second step is O(MK). 2 The time complexity of the third step is O(MK). Where γ i This represents the number of genes in cell i with non-zero values. Since each cell has non-zero expression on only a very limited number of genes, Furthermore, since M and K are often much smaller than N, the total time complexity of the embodiments in this application is O(N+MK). 2 +MK)=O(N). The space complexity is O(N). Therefore, the embodiments of this application can reduce the time complexity and space complexity, thereby improving the computational efficiency.

[0132] The marking feature determination device provided in the embodiments of this application will be described below.

[0133] Reference Figure 9 This application also provides a marker feature determination device, which includes:

[0134] The data acquisition unit 901 is used to acquire target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells;

[0135] The expression level determination unit 902 is used to statistically analyze the average expression level and expression ratio of each candidate feature in different cell types.

[0136] The first inter-class comparison unit 903 is used to determine the first inter-class comparison data for each candidate feature based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type.

[0137] The second inter-class comparison unit 904 is used to determine the second inter-class comparison data for each candidate feature based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type.

[0138] The marker feature determination unit 905 is used to determine the marker features of the target cell type based on the comparison data between the first class and the comparison data between the second class.

[0139] It is evident that the content of the above-described methods for determining the mark features is applicable to the embodiments of this mark feature determining device. The specific functions implemented by this embodiment of the mark feature determining device are the same as those of the above-described methods for determining the mark features, and the beneficial effects achieved are also the same as those achieved by the above-described methods for determining the mark features.

[0140] Reference Figure 10 , Figure 10 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0141] The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0142] The memory 1002 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and is called and executed by the processor 1001 using the flag feature determination method of the embodiments of this application.

[0143] Input / output interface 1003 is used to implement information input and output;

[0144] The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0145] Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004);

[0146] The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0147] This application also provides a computer program product, which includes a computer program. A processor of a computer device reads and executes the computer program, causing the computer device to perform the aforementioned method for determining the marker feature.

[0148] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for determining the marker feature.

[0149] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0150] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0151] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0152] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0153] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0154] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0155] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0156] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0157] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0158] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0159] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0160] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for determining a marker feature, characterized in that, The method includes: Obtain target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells; The average expression level and expression ratio of each candidate feature in different cell types were statistically analyzed. For each candidate feature, inter-class comparison data is determined based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type. For each candidate feature, inter-class comparison data is determined based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type. The marker features of the target cell type are determined based on the first inter-class comparison data and the second inter-class comparison data.

2. The method according to claim 1, characterized in that, For each candidate feature, determining the first-class comparison data based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in non-target cell types includes: The average expression levels of the candidate features in multiple non-target cell types are compared to determine the reference average expression level with the largest value. The average expression difference is calculated by comparing the average expression level of the candidate feature in the target cell type with the reference average expression level. The first inter-class comparison data is determined from the average expression difference and the first preset value.

3. The method according to claim 1, characterized in that, For each candidate feature, determining the second inter-class comparison data based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type includes: The expression ratios of the candidate features in multiple non-target cell types are compared to determine the reference expression ratio with the largest value. The difference between the expression ratio of the candidate feature in the target cell type and the reference expression ratio is calculated to obtain the ratio difference. The second inter-class comparison data is determined from the ratio difference and the second preset value.

4. The method according to claim 1, characterized in that, The step of determining the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data includes: Determine the first comparison threshold and the second comparison threshold; If the first inter-class comparison data is greater than the first comparison threshold, and the second inter-class comparison data is greater than the second comparison threshold, then the candidate feature is used as the marker feature.

5. The method according to claim 4, characterized in that, After determining the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data, the method further includes: Feature relationship data is constructed based on the target cell type, the corresponding marker features of the target cell type, and the candidate features; The first comparison threshold and the second comparison threshold are adjusted based on the feature relationship data.

6. The method according to claim 1, characterized in that, A method for calculating the average expression level of each candidate feature across different cell types includes: For each cell type, the target expression level of the candidate feature in that cell type is calculated. Calculate the total expression level of the candidate features across all cell types; For each cell type, the average expression level of the candidate feature in that cell type is calculated based on the target expression level and the total expression level corresponding to that cell type.

7. The method according to claim 1, characterized in that, A method for calculating the expression ratio of each candidate feature in different cell types includes: For each cell type, the number of first cells in the cell type that express non-zero expression of the candidate feature is counted, and the number of second cells in the cell type is also counted. For each cell type, the expression ratio of the candidate feature in that cell type is calculated based on the number of first cells and the number of second cells corresponding to that cell type.

8. A device for determining a marker feature, characterized in that, The device includes: A data acquisition unit is used to acquire target expression data, wherein the target expression data is used to represent the expression level of each candidate feature in different cells; The expression level determination unit is used to calculate the average expression level and expression ratio of each candidate feature in different cell types. The first inter-class comparison unit is used to determine inter-class comparison data for each candidate feature based on the average expression level of the candidate feature in the target cell type and the average expression level of the candidate feature in the non-target cell type. The second inter-class comparison unit is used to determine the second inter-class comparison data for each candidate feature based on the expression ratio of the candidate feature in the target cell type and the expression ratio of the candidate feature in the non-target cell type. The marker feature determination unit is used to determine the marker features of the target cell type based on the first inter-class comparison data and the second inter-class comparison data.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the marker feature determination method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the marker feature determination method according to any one of claims 1 to 7.

11. A computer program product comprising a computer program that is read and executed by a processor of a computer device, causing the computer device to perform the marking feature determination method according to any one of claims 1 to 7.

Images